aboutsummaryrefslogtreecommitdiffstats
path: root/erts/emulator/drivers/unix
AgeCommit message (Collapse)Author
2019-01-28erts: Add magic port control numbersSverker Eriksson
to increase the probablity of a nice badarg from erlang:port_control.
2018-06-18Update copyright yearHenrik Nord
2017-11-30Reimplement efile_drv as a dirty NIFJohn Högberg
This improves the latency of file operations as dirty schedulers are a bit more eager to run jobs than async threads, and use a single global queue rather than per-thread queues, eliminating the risk of a job stalling behind a long-running job on the same thread while other async threads sit idle. There's no such thing as a free lunch though; the lowered latency comes at the cost of increased busy-waiting which may have an adverse effect on some applications. This behavior can be tweaked with the +sbwt flag, but unfortunately it affects all types of schedulers and not just dirty ones. We plan to add type-specific flags at a later stage. sendfile has been moved to inet_drv to lessen the effect of a nasty race; the cooperation between inet_drv and efile has never been airtight and the socket dying at the wrong time (Regardless of reason) could result in fd aliasing. Moving it to the inet driver makes it impossible to trigger this by closing the socket in the middle of a sendfile operation, while still allowing it to be aborted -- something that can't be done if it stays in the file driver. The race still occurs if the controlling process dies in the short window between dispatching the sendfile operation and the dup(2) call in the driver, but it's much less likely to happen now. A proper fix is in the works. -- Notable functional differences: * The use_threads option for file:sendfile/5 no longer has any effect. * The file-specific DTrace probes have been removed. The same effect can be achieved with normal tracing together with the nif__entry/nif__return probes to track scheduling. -- OTP-14256
2017-11-20Merge branch 'maint'Hans Nilsson
* maint: Updated OTP version Prepare release ssh: testcases for space trailing Hello msg ssh: Don't remove trailing WS in Hello msg ssh: dialyzer fixes ssh: Fix broken error handling during session setup Remove invalid EINTR loop around close(2) Conflicts: OTP_VERSION
2017-11-17Merge branch 'john/erts/fix-close-eintr/OTP-14775' into maint-20Erlang/OTP
* john/erts/fix-close-eintr/OTP-14775: Remove invalid EINTR loop around close(2)
2017-11-13Remove invalid EINTR loop around close(2)John Högberg
Retrying close(2) on anything other than HP-UX is likely to close something entirely different. POSIX says that the state of the file descriptor is unspecified, and Linux/BSD guarantee that it's closed on return.
2017-09-14Merge branch 'maint'Dan Gudmundsson
* maint: stdlib: Fix jumping to beginning or end of line Fix del_chars not considering wide chars and update buffer length before calling write_buf Make cp_pos_to_col function aware of the ANSI escape codes
2017-09-13Fix del_chars not considering wide chars and update buffer length before ↵Glauber Campinho
calling write_buf After deleting the chars the function `del_chars` was considering the code points to move the cursor back and not the graphemes
2017-09-11Merge branch 'lukas/erts/non-smp-removal-cleanup/OTP-14518'Lukas Larsson
* lukas/erts/non-smp-removal-cleanup/OTP-14518: Make estone work with older releases erts: Allow read in ttsl driver to return EAGAIN syntax_tools: Fix makefile dep erts: non-smp removal cleanup in erlexec
2017-09-11erts: Allow read in ttsl driver to return EAGAINLukas Larsson
2017-09-01Make cp_pos_to_col function aware of the ANSI escape codesGlauber Campinho
This fixes the issue with the function `move_cursor` described in #1536 and that causes the bug described in https://github.com/elixir-lang/elixir/issues/6504. The function `cp_pos_to_col` maps the current position in the buffer to the correct column on the screen, but it didn't handle ANSI escape codes. Since the ANSI escape codes aren't visible, the `cp_pos_to_col` now skips them when executing the calculations using `ansi_escape_width`. This new function only considers color escape codes, but also handles invalid codes.
2017-08-21Fix ANSI support in the consoleGlauber Campinho
The ANSI support doesn't work properly with edlin, the issue can be noticed when you try to use the history of the shell and the prompt prefix has ANSI (https://github.com/elixir-lang/elixir/issues/6448). The problem is that when a `\e` character appears, it handles it like a new line, dropping the buffer before it. The solution is to always add the `\e` to the buffer like a regular character and handle it when writing the buffer instead.
2017-08-17Merge branch 'dgud/stdlib/edit-unicode' into maintDan Gudmundsson
* dgud/stdlib/edit-unicode: stdlib: Improve edlin handling of unicode chars OTP-14542
2017-08-16stdlib: Improve edlin handling of unicode charsDan Gudmundsson
Let edlin handle grapheme clusters instead of codepoints to improve the handling multi-codepoints characters. The ttsl driver (and protocol) still expects all lengths as codepoints. Previously it was expected that each codepoint used (at least) one terminal column for each codepoint, and a hack was made for wide characters (multicolumn) by patching in TAGGED characters to occupy the extra space so that codepoint index was equal column index. This didn't work at all for combining codepoints that do not occupy any more space than the previous character. Improved this handling by calculating column positions in move_cursor. This is based on wcwidth() and is not perfect, wcwidth() is wrong for some codepoints and wcwidth() can not know with Hangul graphemes for example. But it works better than before without making a major change in the protocol.
2017-06-07erts: Fix sendfile closeduring scenario on sunosLukas Larsson
On Solaris, giving a too long sfv_len results in an EINVAL error, but data is still transmitted and len is correctly. So we translate this to a success with that amount of data sent. This may hide some other errors that causes EINVAL, but it is the best we can do for now.
2017-05-04Update copyright yearRaimo Niskanen
2017-04-04Merge branch 'maint'Ingela Anderton Andin
Conflicts: OTP_VERSION erts/vsn.mk lib/crypto/c_src/crypto.c lib/crypto/src/crypto.erl lib/ssh/src/ssh.erl
2017-03-29Close FD after trying to open a directoryRaimo Niskanen
2017-02-14Fixed typos in ertsAndrew Dryga
2017-02-07Use fstat if it exists in efile_openfileRaimo Niskanen
2016-04-13Merge branch 'henrik/update-copyrightyear'Henrik Nord
* henrik/update-copyrightyear: update copyright-year
2016-04-07Merge branch 'bjorn/erts/huge-file-fix/OTP-13461'Björn Gustavsson
* bjorn/erts/huge-file-fix/OTP-13461: Handle multi-giga byte writes to files
2016-04-04Handle multi-giga byte writes to filesBjörn Gustavsson
Test cases that write 4Gb to a file at once would fail on OS X and FreeBSD. By running a simple test program on OS X (El Capitan 10.11.4/Darwin 15.4.0), I found that writev() can handle more than 4Gb of data, while write() only can handle less than 2Gb. (Note that efile_drv.c will use write() if there is only one element in the io vector, and writev() if there is more than one.) It is tempting to attempt to piggy-back on the existing mechanism for segmenting write operations in efile_drv.c, but because of the complex code I find it too dangerous, both from a correctness and performance perspective. Instead do the change in unix_efile.c, which is considerably simpler.
2016-03-31Refactor time_t in efile_drvBjörn-Egil Dahlberg
2016-03-30erts: Fix "hanging" VM caused by exiting tty_sl driverSverker Eriksson
Bug introduced on master in a31eab5469b7740d.
2016-03-15Fix ttsl_drv logging without TERMCAPBjörn-Egil Dahlberg
2016-03-15update copyright-yearHenrik Nord
2016-03-15Merge branch 'egil/fix-fdatasync-mac/OTP-13411'Björn-Egil Dahlberg
* egil/fix-fdatasync-mac/OTP-13411: erts: Use fcntl(fd, F_FULLFSYNC) instead of fdatasync on Mac OSX
2016-03-15erts: Use fcntl(fd, F_FULLFSYNC) instead of fdatasync on Mac OSXBjörn-Egil Dahlberg
The syscall fdatasync does not work as intended on Mac OSX. Both the function fsync and fdatasync now uses fcntl(fd, F_FULLFSYNC) on Mac OSX.
2016-03-14erts: Increase ttsl_drv logging capabilitiesBjörn-Egil Dahlberg
2016-01-27Merge branch 'theom/freebsd-sendfile-patch-2/OTP-13271' into maintLukas Larsson
* theom/freebsd-sendfile-patch-2/OTP-13271: erts: Fix sendfile:ing of large files on FreeBSD
2016-01-27erts: Fix sendfile:ing of large files on FreeBSDJP
If the file was larger than the OS send buffer the call would fail before this patch.
2015-09-17Merge branch 'lukas/erts/ttsl_eintr/OTP-12987' into maintRickard Green
* lukas/erts/ttsl_eintr/OTP-12987: erts: Make sure to deal with EINTR write failures
2015-09-16erts: Make sure to deal with EINTR write failuresLukas Larsson
2015-08-27Fix ethread events with timeoutRickard Green
Lots of pthread platforms unnecessarily falled back on the pipe/select solution. This since we tried to use the same monotonic clock source for pthread_cond_timedwait() as used by OS monotonic time. This has been fixed on most platforms by using another clock source. Darwin can however not use pthread_cond_timedwait() with monotonic clock source and has to use the pipe/select solution. On darwin we now use select with _DARWIN_UNLIMITED_SELECT in order to be able to handle a large amount of file descriptors.
2015-08-18erts: Fix binary memory leak in ttsl driverLukas Larsson
2015-06-18Change license text to APLv2Bruce Yinhe
2014-12-19erts: Rename sys_sigset to sys_signalLukas Larsson
Also removed old legacy fallback that is no longer used
2014-10-20Merge branch 'lukas/erts/non-blocking-shell'Lukas Larsson
* lukas/erts/non-blocking-shell: Fix io:columns/0 timeout when invoked via user kernel,ssh: Add synchronous user_drv protocol erts: Make writing to non-tty fds non-blocking erts: Make tty driver non-blocking
2014-10-16kernel,ssh: Add synchronous user_drv protocolLukas Larsson
Added a put_chars_sync to the protocol that can be used to talk to user_drv and made group use it. This is needed in order to guarantee that bytes has been pushed to the tty port when doing something like this: io:format("halting\n"),erlang:halt(0). Before this change the halting message could be lost in the message queue of the user_drv process, this is no longer possible. This commit also fixes ssh_cli as that plugs itself in as a user_drv process. OTP-12240
2014-10-16erts: Make tty driver non-blockingLukas Larsson
Instead of using blocking call to fwrite, the tty driver now uses non-blocking calls to writev and queues any output data that cannot be written into the driver queue. Without this change an stdout write could block an entire scheduler if for some reason the pseudo tty on the other side does not consume the output of the Erlang shell. OTP-12239
2014-08-25Fix misspellings of 'another'Tuncer Ayaz
2014-05-14Fix efile_openfile() to handle stat() failureMikael Pettersson
If the initial stat() fails then efile_openfile() will still proceed to open() the file. If that succeeds and the caller passed a non-NULL pSize, then it will copy bogus data from the statbuf into *pSize. This has been observed to cause file:read_file/1 to return truncated file data with no error indication. The use case involved a large file system mounted via NFS, with some directories containing large number of files, and NFS mount options that allow the NFS client to return EIO if the NFS server does not respond quickly enough. Depending on the caching state of the client and server machines, a few stat() calls (fewer than 1 per 10 million) would take long enough to trigger EIO errors, but subsequent open() calls would succeed, and read_file/1 would return truncated data. This sequence of events has been observed via "strace" on beam.smp. Signed-off-by: Mikael Pettersson <[email protected]>
2014-02-24erts: Fix unix efile assertLukas Larsson
If writev return an error (eg ENOSPC) we do not want to abort here but instead propagate upwards into erlang.
2014-02-24ose: efile driver updates.Jonas Karlsson
2013-11-15Add sync option to file:open/2Joseph Blomstedt
The sync option adds the POSIX O_SYNC flag to the open system call on platforms that support the flag or its equivalent, e.g., FILE_FLAG_WRITE_THROUGH on Windows. For platforms that don't support it, file:open/2 returns {error, enotsup} if the sync option is passed in. The semantics of O_SYNC are platform-specific. For example, not all platforms guarantee that all file metadata are written to the disk along with the file data when the flag is in effect. This issue is noted in the documentation this commit adds for the sync option. Add a test for the sync option. Note however that the underlying OS semantics for O_SYNC can't be tested automatically in any practical way, so the test assumes the OS does the right thing with the flag when present. For manual verification, dtruss on OS X and strace on Linux were both run against beam processes to watch calls to open(), and file:open/2 was called in Erlang shells to open files for writing, both with and without the sync option. Both the dtruss output and the strace output showed that the O_SYNC flag was present in the open() calls when sync was specified and was clear when sync was not specified.
2013-09-12Remove ^L characters hidden randomly in the code. Not those used in text ↵Pierre Fenoll
files as delimiters. While working on a tool that processes Erlang code and testing it against this repo, I found out about those little sneaky 0xff. I thought it may be of help to other people build such tools to remove non-conforming-to-standard characters.
2013-06-12Update copyright yearsBjörn-Egil Dahlberg
2013-05-14Properly guard WIDE_TAG use with HAVE_WCWIDTH in ttsl_drvAnthony Ramine
2013-05-08Merge branch 'nox/wide-chars/OTP-11088' into maintFredrik Gustafsson
* nox/wide-chars/OTP-11088: Support wide characters in the shell through wcwidth() Fix bogus DEBUGLOG() incantations in ttsl_drv